This proposal focuses on short-term, modality-specific memory mechanisms in "specifically language impaired" (SLI) children. There have been few recent investigations of verbal memory in the speech-language pathology literature. There have, however, been recent breakthroughs in understanding the perceptual and conceptual mechanisms underlying memory in normal adults that could be beneficially applied to SLI children. The proposed research will be based on two phenomena of short-term verbal memory. First, the modality effect is the finding that the accuracy of serial verbal recall is greater when stimuli are presented to the auditory than to the visual modality (unlike stimuli in simultaneous spatial arrays, which are better recalled from the visual modality). Second, the suffix effect is the finding that the auditory modality superiority in verbal recall is much reduced when the spoken list is followed by a final spoken item that is not to be recalled. These two phenomena have been extensively investigated and appear to indicate that normal subjects make use of temporary acoustic and phonetic-memory codes that are especially well-suited for the retention of serial verbal information, although they are vulnerable to interference. Recent studies also demonstrate that the modality and suffix effects are obtained in the memory and comprehension of natural, linguistically coherent language. In the proposed experiments, modality and suffix effects will be examined in SLI children and in other groups matched for age and reading abilities, in order to determine if the use of auditory modality-specific memory or its vulnerability to interference differs in these groups. List recall procedures will be used in the early experiments. Later studies extend the findings to the comprehension of coherent language.